CHAPTER 8

 

Sovereign Debt Restructuring

A COASEAN PERSPECTIVE

James A. Haley

In the words of the great American social philosopher Yogi Berra, “It’s déjà vu all over again” with respect to sovereign debt restructuring: in late 2014, Mexico successfully issued New York law bonds incorporating revised collective action clauses (CACs).¹ The new clauses, which strengthen existing provisions to assuage creditor coordination problems and promote the timely, orderly restructuring of sovereign debt, were developed by the International Capital Market Association (ICMA) in collaboration with the U.S. Treasury, the International Monetary Fund (IMF), and the Institute for International Finance.

A little more than a decade ago, Mexico likewise demonstrated leadership in introducing first-generation CACs in its New York–law bonds. At the time, CACs were viewed as the cornerstone of the “voluntary” approach to sovereign debt restructuring in the debate between backers of a market-led approach to address creditor coordination problems and advocates of a formal statutory approach incorporating elements of domestic bankruptcy frameworks.² Mexico’s introduction of CACs in February 2003, together with the recognition that statutory approaches lacked the political support needed to be implemented, effectively ended the debate and relegated the statutory approach to the proverbial back burner.³

Significant progress was made in terms of sovereign debt restructuring in the years between the initial Mexican bond issue with CACs and today. Soon after Mexico’s pathbreaking 2003 issue, three-quarters of all new sovereign bond issues included CACs; by 2010, more than 90 percent of all new issues included these clauses. Moreover, through much of the past decade, sovereign debt restructurings were completed with relative speed and were notable for the absence of litigation.⁴ Indeed, the experience was such that by late 2011 supporters of the contractual approach could claim vindication. If not the “best of all possible worlds” of Voltaire’s Dr. Pangloss, the process for restructurings seemed at least to be working; then came Greece and Argentina.

By early 2012, the sovereign debt restructuring landscape had changed. With more than 90 percent of its debt stock governed by domestic law, Greece retroactively inserted an aggregated collective action mechanism across its bonds with voting thresholds very favorable to the debtor, much to the chagrin of its bondholders. Meanwhile, holdout creditors of Argentina who chose not to participate in earlier restructurings secured a favorable judgment upholding a pari passu clause in the 1994 Fiscal Agency Agreement that forbids Argentina to pay other debts unless it also pays the holdout creditor on a pro rata basis, which in the court order providing injunctive relief is interpreted as “ratable” payments.⁵

These “new challenges” have resurrected “old debates” about the merits of contractual versus statutory approaches to sovereign debt restructuring.⁶ And Anna Gelpern, a longtime, self-proclaimed skeptic of bankruptcy for sovereigns, surveying recent debt restructuring experiences, has concluded: “The existing system for reducing sovereign debt is deeply dysfunctional and produces bad law.”⁷ The arbitrary nature of retroactive insertion of CACs in Greek debt and the use of pari passu against investors that accepted restructured bonds have tarnished the Panglossian perspective on sovereign debt restructurings that prevailed just a few years ago and have reenergized efforts to facilitate more predictable outcomes for creditors and debtors and fairer treatment across creditors. The ICMA model clauses on which Mexico’s late-2014 bond is based are a response to these concerns of predictability and fairness.

As a preview of what follows, I make three (admittedly unoriginal) points.

First, the new clauses are a useful and potentially important instrument for dealing with the problem of holdout creditors. Indeed, given that they address key weaknesses and obvious problems in existing bonding “technology,” the ICMA clauses are undoubtedly a “big deal.”⁸

Second, the new clauses are not a panacea. This assessment reflects the fact that it will take some time for these clauses to be embedded in the stock of outstanding bonds. The benefits from them, meanwhile, can be expected to dissipate over time as contractual practices evolve. Moreover, whatever their merits, the new clauses do not address the issues of unenforceability and discharge of sovereign debts.

Third, the debate between voluntary/contractual and statutory approaches is a false dichotomy. Contractual approaches will necessarily be incomplete and the design of “institutions,” whether bankruptcy provisions embodied in formal treaty or the responses of existing international financial institutions, will influence the outcome of sovereign debt restructurings.

The first of these points is factual and, I think, noncontroversial. The latter two are more conjectural in nature and subject to possible debate.

IMPACT OF NEW CLAUSES

The ICMA model clauses adopted by Mexico and other issuers in late 2014 are intended to make it harder for holdout creditors to disrupt future sovereign debt restructurings by addressing the aggregation problem under the series-by-series voting process that has been incorporated in CACs since 2003. In series-by-series voting, a small minority of creditors controlling a single bond issue can impede a restructuring acceptable to the broad majority of investors holding other series. The problem is that creditors will be reluctant to reduce their claims on distressed sovereigns if other creditors are not also participating in the restructuring. As a result, if a small minority of investors in a single bond issue refuses to accept modified terms, they could potentially block the entire restructuring. Rather than risk failure, sovereign debtors may pay the minority creditors in full. But paying off the minority creates an incentive for holdout investors to “free ride” on the majority of creditors that are prepared to reduce their claims.⁹

To make it harder for holdout creditors to disrupt future bond restructurings, the ICMA model clauses allow issuers to organize a vote across bond series in two ways, in addition to the existing series-by-series mechanism.¹⁰ The first method is a “two-limb” voting process in which the threshold for single series approval is reduced from 75 to 50 percent, subject to approval by two-thirds of the entire pool of bondholders. The second modification to voting proposed by the ICMA model clauses is “single-limb” voting by all bondholders, across all series, with a 75 percent threshold for amending terms. Given the potential for abuse by minority creditors, the single-limb rule requires the issuer to offer the same menu of participation options to all creditors—the “uniform consideration” clause.¹¹

The power of the proposed changes comes from the reduction in the ability of minority investors to block restructurings. The change from series-by-series voting to the single-limb mechanism is striking, particularly in cases in which the sovereign has a large number of outstanding bond issues, some with relatively small nominal amounts outstanding. Under such conditions, blocking the restructuring of a single series in a series-by-series vote requires a relatively modest investment. In contrast, the single-limb approach requires a much larger investment, equivalent to 25.1 percent of the entire outstanding debt stock. For a large issuer with a large stock of debt, this could represent an insurmountable obstacle. At the same time, a 75 percent threshold is sufficiently high that successful restructurings will continue to be difficult to achieve, requiring the cooperation of the broad majority of creditors. In this regard, it is interesting to note the incentive that the single-limb mechanism provides for the debtor to work with supportive creditors in designing a voting strategy that maximizes the likelihood of success to the benefit of most creditors.

The ICMA clauses also refine the troublesome language of pari passu clauses that, in the hands of holdout creditors and following a favorable judicial ruling in New York, has led Argentina into selective default and thrust the issue of sovereign debt restructuring once more onto the international policy agenda. By neutralizing the notion of ratable payments, the new clause should reduce the risk of copycat litigation. There is residual uncertainty regarding contractual interpretation, however, and only time will tell whether the new clause will achieve its objective.

MARKET EVOLUTION AND BONDING TECHNOLOGY

Notwithstanding the potential benefits from ICMA model clauses, the environment for sovereign debt restructuring will likely remain challenging.¹² In the first instance, the benefits from aggregation mechanisms will take time to accrue owing to the stock/flow issue: the IMF estimates that roughly 30 percent of the US$900 billion in outstanding bonds globally will mature in more than 10 years.¹³ The potential for holdouts will clearly remain for some time to come. In the interim, there is a risk that the partial introduction of mechanisms to facilitate restructuring could increase the risk of free riding, as ease of restructuring bond issues with ICMA clauses increases the potential returns from disrupting restructurings.¹⁴

More generally, there are residual challenges to a purely contractual framework for debt restructuring coming from a number of sources. The first is interpretation uncertainty. The new ICMA model clauses will be subject to judicial interpretation; as demonstrated with respect to pari passu in the case of Argentina, such rulings may not adhere to prevailing opinion. Over time, the meaning and interpretation of new clauses will be clarified, defined, and incorporated in jurisprudence, reducing the degree of uncertainty. But there will be uncertainty that will affect behavior of debtors and creditors alike.

In addition, regardless of their merits, the new clauses do not address the basic issues of unenforceability and debt discharge that plague sovereign debt restructurings.¹⁵ In contrast to domestic commercial bankruptcies, sovereigns suffering from excessive debt cannot secure a judicial discharge of debts that binds all creditors and prevents disruptive litigation and asset stripping. And apart from nuisance attempts to seize assets, there is very little that creditors can do to force payment from sovereigns that do not pay their debts. As a result, it is reasonable to assume that sovereign debt contracts will have a self-enforcing element. While better contract design can assist a sovereign seeking a cooperative solution that preserves asset values for creditors and restores the economy to growth, there will be residual uncertainty concerning “ability” and “willingness” to honor existing contracts.¹⁶

One implication of this uncertainty is that, short of some internationally binding agreement that enforces the use of standard, nonvarying terms, contractual approaches will be subject to innovation and evolution as market conditions and the needs of sovereign borrowers change. There may well be circumstances in which sovereign borrowers and potential investors develop innovations that immunize an issue against a particular formulation of an aggregation clause.¹⁷ Moreover, the introduction of “near-debt” instruments, which share some equity features, may be helpful in reducing the vulnerability of a sovereign’s balance sheet in response to adverse shocks within a certain range, but could complicate restructuring in the event of a much larger shock. Finally, the potential for governments to legislate institutional barriers to efficient recontracting cannot be discounted.

A FALSE DEBATE?

The upshot of this dynamic view of market evolution and contract design is that the benefits of a particular contract innovation are likely to erode over time.¹⁸ Although better contracts should deliver better outcomes in terms of lower deadweight losses to creditors and sovereign borrowers, the size of those benefits will be subject to uncertainty.

In theory, we should be indifferent with respect to approaches if they result in the same outcome in equilibrium—defined such that neither the creditors nor the borrower seeks to modify the payment schedule (mutually agreed to in the case of voluntary outcomes or imposed in the case of statutory bankruptcy frameworks). This might be the case, say, if the obligations under debt contracts are articulated clearly and resulting rights over the sovereign’s tax revenues so defined are effectively enforced.

In this regard, advocates of voluntary approaches implicitly invoke the so-called property rights school to support their case. In his pioneering paper, The Problem of Social Costs, Ronald Coase argued that, in the absence of transactions costs, bargaining between private parties will ensure the efficient allocation of resources, irrespective of the initial distribution of ownership rights.¹⁹ In practice, however, as Arrow (1979) demonstrated, this proposition is equivalent to saying that the outcome is Pareto-optimal.²⁰

This result depends critically on strict assumptions with respect to the information sets of the different players. Specifically, that every player knows every other player’s payoff as a function of the strategies played—the offers and counteroffers that the other side will make in response to their offers. The failure of the underlying postulates of the bargaining approach can result in a situation in which the two sides get stuck in a Pareto-dominated point. The problem is that rather than converge on a stable equilibrium, successive iterations of offers and counteroffers are based on misperceptions of what the other side is prepared to accept. This could account for protracted sovereign debt restructurings in which economic losses grow and asset values shrink, ultimately resulting in restructurings that in the words of the IMF, are “too little, too late.”²¹

More fundamentally, in contrast to the basic assumption of the property rights school, resource transfers from the debtor to the creditor are not well defined in the case of sovereign debt: there may well be ex ante agreement on payments between creditor and debtor, but there is weak enforcement ex post; indeed, in many jurisdictions, unenforceability is enshrined in law to promote the comity of nations.²² Meanwhile, the “unintelligibility” of the legal rules governing sovereign restructuring, as one legal scholar characterizes the status quo, creates uncertainty that amounts to a transactions cost, violating the postulates of the Coase theorem.²³

This discussion underscores a false dichotomy between “contractual” and “statutory” approaches and rightly emphasizes the issues of enforceability and limitation of contract design—practical problems that the property rights school assumes away by recourse to zero transactions costs and full information. There is reason to believe that Coase would approve of this focus on enforceability and contract design. Nearly thirty years after its publication, he feared his seminal paper had been misinterpreted.

He wrote, “Its influence on economic analysis has been less beneficial than I had hoped.” His aim, he said, was not simply to describe what life would be in a world without transaction costs, but rather, “to make clear the role which transaction costs do, and should, play in the fashioning of the institutions which make up the economic system.” What has become known since as the “Coasean World,”—where rational actors transact freely without need for institutions, firms, or even law—“is really the world of modern economic theory, one which I was hoping to persuade economists to leave.”²⁴

A Coasean perspective on sovereign debt restructuring recognizes that institutions, whether bankruptcy provisions embodied in formal treaty or promulgated through a “soft law” approach in the policies of existing international financial institutions, will influence the outcome of debt restructurings.²⁵ These policies jointly affect the pace and progress of the process. For example, strict access limits on the size of IMF support packages with clearly defined conditions for exceptional access could influence negotiations between sovereign borrowers and their creditors by anchoring expectations of IMF assistance in the event of severe balance of payment difficulties.²⁶ But absent some means to contain the fallout from limiting financing, attempts to impose strengthened access limits and constrain discretion may not be credible and would not, therefore, affect behavior. Similarly, an internationally recognized stay on litigation (or “standstill”) sanctioned by the IMF or some other international body could allow for the orderly restructuring of claims and provide the sovereign time to introduce policies that “grow the pie” to the benefit of domestic citizens and foreign creditors alike. And IMF lending into arrears, which provides access to financing to avoid a draconian compression of imports, can be thought of as the analogue of debtor-in-possession financing for sovereigns.²⁷ IMF lending has potential external effects, however, in that its preferred creditor status can subordinate private claims and complicate debt restructurings, particularly if there is a nonnegligible probability that further restructuring may be required in the future.

In this world, the contractual approach and efforts to integrate key elements of domestic bankruptcy in the framework for sovereign debt restructuring are not mutually exclusive; in fact, they complement each other. This is observed at the domestic level, where bargaining occurs and voluntary debt restructurings are completed “in the shadow of the courthouse,” because creditors and borrowers know the consequences of failure. The threat of an involuntary solution “through the courthouse” creates an incentive to do the deal privately, while well-developed legal rules reduce the uncertainty attached to potential outcomes and guide voluntary settlements.²⁸ The sovereign debt analogue, perhaps, is the risk that failure to reach agreement on a debt restructuring in the context of an IMF-supported program could lead to an economic collapse detrimental to all parties.

This Coasean perspective is also aligned with the agency costs approach to understanding the problems of sovereign debt and the role of the IMF.²⁹ In this approach, the IMF can act as a bonding mechanism to assist countries’ access international capital markets by virtue of its superior monitoring ability derived from its ongoing policy dialogue and confidential discussions with national authorities. Committing to IMF-supported programs allows sovereign borrowers to benefit from higher levels of borrowing and on more favorable terms than would be possible without the IMF. The outstanding question, though, is whether there are additional public policy interventions that, in conjunction with traditional IMF support, could facilitate better outcomes should it be necessary to recontract a debt.³⁰

So, where do we stand with respect to the contractual approach to sovereign debt restructuring?

We should, clearly, welcome the potential benefits from ICMA model clauses. The introduction of such clauses should mitigate some of the problems arising from the new challenges posed by the Greek restructuring and the litigation against Argentina. But it is both too early and too risky to declare that these clauses represent the best of all possible worlds. In the presence of weak enforcement and absent debt discharge, contractual approaches will necessarily be incomplete. Because sovereign debt contracts are incomplete, they will need to be restructured in the face, say, of severe negative shocks to output. Clear rules of the game—whether in formal statutory frameworks or in the policies adopted by the institutions of international cooperation—would provide guidance to this process and facilitate timely, orderly restructurings. In this respect, we need to understand the institutional environment in which these clauses are used and, where appropriate, make necessary changes to institutional practice, even if the political will does not exist to take more fundamental steps.

NOTES

Helpful comments from Martin Guzman, Paul Jenkins, Trevor Lessard, and Nicholas Marion are gratefully acknowledged. The usual caveat applies.

 

1. The Mexican bond was not the first to incorporate the revised clauses. A few days earlier Vietnam had made a private placement incorporating the ICMA clauses; Kazakhstan had also issued bonds that include most elements of the new terms prior to Mexico.

2. The apotheosis of the statutory approach was the proposal for a Sovereign Debt Restructuring Mechanism developed by IMF staff under the direction of then first deputy managing director Anne Krueger.

3. See the discussion in Ocampo (2016).

4. See IMF (2013).

5. The court’s interpretation is such that, if Argentina pays 100 percent of the amounts owing on restructured bonds, it must pay 100 percent of the amounts owing on the un-restructured bonds held by the holdouts.

6. Haley (2014).

7. Gelpern (2013b).

8. Gelpern (2014).

9. In the March 2012 Greek restructuring, foreign (U.K.) law bonds were not restructured and were paid in full.

10. An excellent discussion of these issues is found in Makoff and Kahn (2015).

11. In effect, this feature replicates provisions in U.S. bankruptcy law under which all bonds are accelerated to par in the event of bankruptcy, putting all bondholders on an equal footing in the negotiation process.

12. Proposals for the further evolution of the contractual approach include an automatic standstill provision and creditor engagement clauses, requiring sovereign issuers to convene and negotiate with representative creditor committees as a condition of “good faith” negotiations. For a discussion of the potential benefits and drawbacks of creditor committees, see DeSieno (2016).

13. IMF (2014).

14. See Makoff and Kahn (2014). Pitchford and Wright (2010) have suggested, for example, that collective action clauses and other measures to facilitate timely, orderly restructurings increase the returns to holding out intended to extract greater returns from the sovereign.

15. Gelpern (2013a).

16. It is curious that appeals are made to the “sanctity of contracts” in this context. Contracts between private parties are amended under domestic bankruptcy laws, leaving the debtor in possession of the firm’s assets when there is a public policy interest to do so and legal tests are met; it is unclear why a contract between a private creditor and sovereign government (representing social interests) should be inviolate.

17. The risk of such innovations is exacerbated by the problem of excessive debt accumulation arising from the absence of a clear priority of claim in sovereign debt. Bolton and Jeanne (2007), meanwhile, argue that sovereign lenders can increase the likelihood of repayment in the presence of weak contract enforcement by making their claims more difficult to restructure. While such contracts increase the debt capacity of the sovereign borrower in good times, they result in higher deadweight losses in bad states of the world in which a recontracting is required. More recently, Halonen-Akatwijuka and Hart (2013) have contended that parties may deliberately write incomplete contracts (or contracts that are less complete than is feasible). The intuition behind their result is that the specification of remedies in the event of a particular event may increase influence expectations with respect to events not subject to state-contingent clauses and lead to higher deadweight losses.

18. As an example, a legal ruling in London in July 2012 could potentially limit the use of exit consents—an innovative legal strategy that had facilitated many restructurings over the previous decade. Under exit consents, bondholders agreeing to a restructuring through a debt exchange simultaneously “consent” to amendments in the nonfinancial terms of the old bonds from which they are “exiting.” These changes reduce the underlying value of the old bonds. Since changes to nonfinancial terms typically require a lower voting threshold, the use of exit consents had thus been an effective means to reduce the incentive to holdout. London courts struck down the use of exit consents on the grounds that they violated English law and the terms of the trust deed of the bonds under litigation. The judgment supported the claims of holdout investors that the clauses in question amounted to an abuse of power by the majority and were “oppressive and unfair” to the minority.

19. Coase (1960).

20. Arrow (1979). Pareto optimality implies that neither party can be made better off without putting the other in an inferior position. The proof is the “folk theorem” of cooperative games: both sides will continue to bargain and strike side agreements until there is no possible outcome that is preferred by either.

21. IMF (2013).

22. See, e.g., the U.S. Foreign Immunities Act.

23. See Gelpern (2013b). I am indebted to an anonymous reviewer for this point.

24. Extract from the University of Chicago Law School, “The Problem of Social Cost,” Coase in memoriam, www.law.uchicago.edu/lawecon/coaseinmemoriam/problemofsocialcost.

25. For a discussion of the “soft law” approach, see Guzman and Stiglitz (2016).

26. See the discussion in Brooks and Lombardi (2016).

27. Under most domestic bankruptcy regimes, a court-sanctioned stay on proceedings and DIP financing reduces the costs of damaging creditor runs. Existing creditors benefit from the breathing space provided to the debtor to reorganize and propose an orderly restructuring to the benefit of all creditors; similarly, while new lending under DIP financing enjoys priority, it helps preserve the asset values of all creditors by, e.g., keeping the firm in operation, preserving the capital of the firm as a “going concern,” and allowing for the introduction of measures to return the operation to profitability. Protection from litigation benefits creditors by preventing the rush to the courthouse and the dissipation of asset values through a disruptive liquidation of assets under fire sale prices.

28. A sovereign debt forum—a venue for the sharing of information and a repository of sovereign debt restructuring “best practices”—could reduce uncertainty and facilitate timely debt restructurings by reducing information asymmetries and building trust. See Gitlin and House (2016).

29. Tirole (2002).

30. One possibility, in addition to calls for more work on mechanisms to facilitate restructuring, might be to provide guarantees on restructured debts to reduce the potential costs associated with serial restructuring. Uncertainty associated with the threat of future restructurings could be dissipated by breaking claims into two components. The first component would be a non-state-contingent element based on very robust assumptions (to avoid a “balanced on a knife-edge” scenario), which would benefit from the guarantee. The second component would entail a state-contingent element providing upside returns to creditors in the event of stronger-than-expected growth owing to favorable shocks, with the IMF performing monitoring and third-party verification in accordance with its role in addressing agency problems.

REFERENCES

Arrow, Kenneth J. 1979. “The Property Rights Doctrine and Demand Revelation Under Incomplete Information.” In Economics and Human Welfare, ed. M. Boskin. New York, Academic Press: 23–39.

Bolton, Patrick, and Olivier Jeanne. August 2007. “Structuring and Restructuring Sovereign Debt: The Role of Bankruptcy Regime.” International Monetary Fund Working Paper WP/07/192, Washington, D.C.

Brooks, Skylar, and Domenico Lombardi. 2016. “Private Creditor Power and the Politics of Sovereign Debt Governance.” In Too Little, Too Late: The Quest to Resolve Sovereign Debt Crises, ed. Martin Guzman, José Antonio Ocampo, and Joseph E. Stiglitz, chapter 3. New York: Columbia University Press.

Coase, Ronald. 1960. “The Problem of Social Costs.” Journal of Law and Economics 3 (October): 1–44.

De Sieno, Timothy. 2016. “Creditor Committees in Sovereign Debt Restructurings: Understanding the Benefits and Addressing Concerns.” In Too Little, Too Late: The Quest to Resolve Sovereign Debt Crises, ed. Martin Guzman, José Antonio Ocampo, and Joseph E. Stiglitz, chapter 9. New York: Columbia University Press.

Gelpern, Anna. 2013a. “Contract Hope and Sovereign Redemption.” Capital Markets Law Journal 8 (no. 2): 132–148.

——. 2013b. “A Skeptic’s Case for Sovereign Bankruptcy.” Houston Law Review 50: 1095–1127.

——. 2014. “ICMA CACs v. 2.0: Mexico Moves in New York.” Credit Slips, November 10.

Guzman, Martin, and Joseph E. Stiglitz. 2016. “Creating a Framework for Sovereign Debt Restructuring That Works.” In Too Little, Too Late: The Quest to Resolve Sovereign Debt Crises, ed. Martin Guzman, José Antonio Ocampo, and Joseph E. Stiglitz, chapter 1. New York: Columbia University Press.

Haley, James A. May 2014. “Sovereign Debt Restructuring: Old Debates, New Challenges.” CIGI Papers No. 32, Waterloo, Ont.

Haley, James A., Halonen-Akatwijuka, Maija, and Oliver D. Hart. April 2013. “More Is Less: Why Parties May Deliberately Write Incomplete Contracts.” NBER Working Paper 19001, Cambridge, Mass.

International Monetary Fund. 2013, April 26. “Sovereign Debt Restructuring—Recent Developments and Implications for the Fund’s Legal and Policy Framework.” Washington, D.C.: https://www.imf.org/external/np/pp/eng/2013/042613.pdf.

——. October 2014. “Strengthening the Contractual Framework to Address Collective Action Problems in Sovereign Debt Restructuring.” Washington, D.C.: https://www.imf.org/external/np/pp/eng/2014/090214.pdf.

Makoff, Gregory, and Robert Kahn. Sovereign Bond Contract Reform: Implementing the ICMA Pari Passu and Collective Action Clauses, CIGI paper no. 56, February 6, 2015: https://www.cigionline.org/publications/sovereign-bond-contract-reform-implementing-new-icma-pari-passu-and-collective-action-c.

Ocampo, José Antonio. 2016. “A Brief History of Sovereign Debt Resolution and a Proposal for a Multilateral Instrument.” In Too Little, Too Late: The Quest to Resolve Sovereign Debt Crises, ed. Martin Guzman, José Antonio Ocampo, and Joseph E. Stiglitz, chapter 10. New York: Columbia University Press.

Pitchford, Rohan, and Mark L. J. Wright. December 2010. “Holdout in Sovereign Debt Restructuring: A Theory of Negotiation in a Weak Contract Environment.” NBER Working Paper 16632, Cambridge, Mass.

Tirole, Jean. 2002. Financial Crises, Liquidity Provision and the International Monetary System. Princeton, N.J.: Princeton University Press.